Automation of Sensor Control in Uninhabited Aerial Vehicles

Abstract : The use of automation in uninhabited aerial vehicles (UAVs) is being explored using simulation to develop a better balance of system autonomy with human interaction for optimal human-system performance. Integrating fixed levels of automation into elements of UAV control has resulted in improved performance of some tasks, but less so when the automation does not fit appropriately to changing UAV tasks and environments. As a consequence, recent research has focused on automation that can be changed during UAV operation. The aim of this study was to compare the performance of operators for some UAV tasks of interest under automation that could be changed (adaptable), fixed automation, and no automation. A secondary aim was to examine the use of automation and its effect on the performance of these tasks in different area densities. Thirty eight participants from the Defence Science and Technology Organisation (DSTO) Edinburgh controlled the sensor of an UAV and undertook typical tasks in a maritime environment. Sensor control was manipulated to be fully automated (fixed), completely manual, or adjustable by the participant between automated and manual modes (adaptable). The results showed that fixed sensor control produced superior performance compared to manual control across tasks, but adaptable control produced additional time improvements compared to fixed control. The benefits of adaptable automation of sensor control and future research directions are discussed.